brian j. irwin...the unnatural history of the sea.) ! index of relative abundance ! multiple visits...
TRANSCRIPT
Brian J. Irwin
Northeast Climate Science Center's Regional Science
Meeting University of Massachusetts, Amherst
May 2017
Georgia Coopera+ve Fish and Wildlife Research Unit Cooperators
Great Lakes Science Center
§ PhD student: Tiffany Vidal (Univ. of Georgia) — MA Division of Marine Fisheries — Understanding the role of variability in fish
population response to changing environmental conditions
§ Co-PIs: Jim Bence (Michigan State Univ.),
Ty Wagner (PA Coop. Unit)
§ Partners: Jim Hoyle (OMNR), Randy Jackson (Cornell Univ.), Chuck Madenjian (USGS), Lars Rudstam (Cornell Univ.)
1) Case-study Example — Long-term monitoring data — Large-scale Ecological Change — Variance partitioning — Is variance structure responsive to perturbation?
2) Implications — Monitoring & Management — Climate Change
“any deviation, or displacement, from the ‘nominal state’ in structure or function at any level of organization. The nominal state is the normal operating range, including expected variance.”
– Odum et al. 1979
“[I]t is not true that a species thus attacked comes back. The disturbed balance often gives a new species ascendancy and destroys forever the old relationships.”
– John Steinbeck. 1951. The Log from the Sea of Cortez. (As cited in Roberts. 2007. The Unnatural History of the Sea.)
§ Index of relative abundance § Multiple visits
- Sites - Years
§ Variable over space and time - Catch - Effort
MI Sea Grant
Credit: NOAA GLERL MI Sea Grant
§ Long-term gillnet surveys — 15 sites, >50 years
1958 early 1990s 2010
Year
1960 1970 1980 1990 2000 2010
Cat
ch /
Net
0
50
100
150
200
250 AverageObserved
Year
1960 1970 1980 1990 2000 2010
Cat
ch /
Net
0
50
100
150
200
250 Average
Time
Eco
syst
em re
spon
se (A) (C)(B)
Spatial Coherent temporal
Ephemeral temporal
Time
Modified from Irwin et al. 2013
𝑏↓𝑗𝑝 ~𝑁(0,𝜎↓𝑏↓𝑝 ↑2 ) Spatial Coherent
𝑌↓𝑡𝑗 ~𝑁𝐵(𝜇↓𝑡𝑗 , 𝜅↓𝑝 ) 𝜇↓𝑡𝑗 = 𝑒↑(𝜂↓𝑡𝑗 )
t = year 𝑗=site 𝒑 = period
𝜂↓𝑡𝑗 = 𝜈↓𝑝 +𝜆(𝑡)+ 𝑎↓𝑡𝑝 + 𝑏↓𝑗𝑝
𝑎↓𝑡𝑝 ~𝑁(0,𝜎↓𝑎↓𝑝 ↑2 )
𝑣𝑎𝑟↓𝑡𝑗 = 𝜇↓𝑡𝑗 + 𝜇↓𝑡𝑗↑2 /𝜅↓𝑝
Vidal et al. 2017
Vidal et al. 2017
Vidal et al. 2017
Vidal et al. 2017
§ Outputs / Results — Capacity building via graduate education — Flexible & transferable modeling approach
• Quantify population responses to large-scale change • More than just testing for mean response (e.g., spatial
homogenization)
§ Collaborations & Stakeholders — Partnerships: Universities (Cornell, UGA), State
(NY DEC, GA DNR), and Federal (NECSC, USGS)
• Question-driven monitoring, commitment to monitoring • Effective working relationships, discussions: questions,
data usage & publication
§ Decision Making — Spatial and temporal population structure has
management & monitoring implications • e.g., priority locations (which sites to preserve?) • e.g., within year or among year changes? • e.g., eliminating sites or skipping years?
§ Gaps — Confronting hypotheses with data
• How do disturbances cause system instability? • “Other” species • “Exploration” monitoring